The isolation of CHO cells with a site conferring a high and reproducible transgene amplification rate

Abstract

Co-amplification of transgenes using the dihydrofolate reductase/methotrexate (DHFR/MTX) system is a widely used method for the isolation of Chinese hamster ovary (CHO) cell lines that secrete high levels of recombinant proteins. A bottleneck in this process is the stepwise selection for MTX resistant populations; which can be slow, tedious and erratic. We sought to speed up and regularize this process by isolating dhfr⁻ CHO cell lines capable of integrating a transgene of interest into a defined chromosomal location that supports a high rate of gene amplification. We isolated 100 independent transfectants carrying a gene for human adenosine deaminase (ada) linked to a φC31 attP site and a portion of the dihydrofolate reductase (dhfr) gene. Measurement of the ada amplification rate in each transfectant using Luria–Delbruck fluctuation analysis revealed a wide clonal variation; sub-cloning showed these rates to be heritable. Site directed recombination was used to insert a transgene carrying a reporter gene for secreted embryonic alkaline phosphatase (SEAP) as well as the remainder of the dhfr gene into the attP site at this location in several of these clones. Subsequent selection for gene amplification of the reconstructed dhfr gene in a high ada amplification candidate clone (DG44-HA-4) yielded reproducible rates of seap gene amplification and concomitant increased levels of SEAP secretion. In contrast, random integrations of the dhfr gene into clone HA-4 did not yield these high levels of amplification. This cell line as well as this method of screening for high amplification rates may prove helpful for the reliable amplification of recombinant genes for therapeutically or diagnostically useful proteins.

Introduction

The phenomenon of gene amplification is often exploited to produce therapeutic recombinant proteins, monoclonal antibodies being the prime example. A gene or genes coding for the therapeutic protein(s) of choice (e.g., monoclonal antibody heavy and light chain) are transfected into a host cell along with a selectable drug-resistance marker gene. After an initial isolation of transfectants exhibiting a minimal level of drug resistance, selection for transfectants that have acquired a modestly higher level of resistance by virtue of amplification of the marker gene is carried out. Many iterations of this selection process can result in cell clones with a high level of resistance, a high level of marker gene expression, and a high number of copies of the marker gene. Since the size of the amplified region is much larger than a single gene (Hamlin, 1992) and since transfected genes usually co-integrate in the host genome, even when co-transfected on separate plasmids (Chen and Chasin, 1998), the gene of interest is also co-amplified and its protein produced at a high level (Wigler et al., 1980).

A popular system for this approach uses a gene specifying the enzyme dihydrofolate reductase (DHFR) as the selective marker (Ringold et al., 1981), methotrexate (MTX) as the drug and a Chinese hamster ovary (CHO) cell line deficient in this ubiquitous enzyme (Urlaub and Chasin, 1980, Urlaub et al., 1983, Jayapal et al., 2007) as a host. CHO cells are capable of gene amplification, a trait not shared by normal cells (Livingstone et al., 1992, Wahl et al., 1984a, Yin et al., 1992). In this case the target of the drug is DHFR, and resistance is often gained by overproducing it. Transfected cells are cultured in a medium lacking a source of purine and thymidine nucleotides. Since DHFR-deficient host cells are unable to synthesize these metabolites only the transfectants can grow. MTX is a specific and tight-binding inhibitor of DHFR, but amplified cells that have undergone dhfr amplification (“amplificants”) can overcome a judiciously chosen concentration of the drug.

Although the dhfr/MTX gene amplification method can result in as much as a 1000-fold increase in gene copy number (Kaufman and Sharp, 1982), it suffers from long and variable development times. Each amplification step brings about only a modest increase in gene copy number, thus severely limiting the concentration of MTX that can be applied at each step. It is common to take six months or longer to isolate a cell line with the desired recombinant protein production level. This time bottleneck inhibits the rapid testing of multiple new candidates for pre-clinical evaluation and ultimately limits how quickly a new drug candidate gets to market (Trill et al., 1995). Use of MTX amplification may also have unforeseen effects on CHO cells since chromosomal translocations often accompany gene amplification. Additionally, amplified genes are not always stable, resulting in a decrease in gene copy number unless selective pressure is maintained. Due to these drawbacks, this approach is often not used despite its potential to yield high gene copy number clones. Industry has reacted by using novel promoter-enhancer and chromatin opening elements and developing high-throughput techniques to isolate rare high producing clones. Despite these drawbacks, regularizing this process and decreasing the time to implement it could lead to its more widespread use in the development of useful biologics.

Although the factors influencing gene amplification are not well understood, it has been shown that clones within a CHO transfectant population exhibit a wide distribution in the yield of amplificants (Kito et al., 2002, Wahl et al., 1984b). The simplest explanation for this variability is a position effect within the CHO genome. We reasoned that the gene amplification regimen could be both shortened and made more reliable by inserting the gene of interest along with the dhfr gene at a defined locus in the CHO genome chosen for supporting a high amplification rate. We identified such sites by screening 100 transfectants for their rate of amplification (as opposed to the frequency of amplified cells) using Luria–Delbruck fluctuation analysis. A chosen site could subsequently be used to insert genes of interest via an included site specific recombination sequence. A winning clone was shown to amplify a dhfr gene along with a gene of interest at a rate higher than randomly integrated transfectants and to secrete a protein of interest at increased levels. This clone or similarly isolated clones may prove advantageous for those using gene amplification in CHO cells to drive high level production of therapeutically or diagnostically useful proteins.